Outro

Throughout this book, I’ve told a story of how one influential strand of causal inference emerged from an unlikely romance between Princeton’s Industrial Relations Section and Harvard’s Statistics Department. This isn’t just academic gossip—it’s my way of understanding why this particular tradition feels different from traditional econometrics, and why methods like instrumental variables, diff-in-diff, RDD, and synthetic control all seem to share a common DNA. When Princeton’s “shoe leather” tradition met Harvard’s statistical rigor, they created something that became hugely influential in applied economics, and then increasingly throughout all of the social sciences, which I’ve tried to capture in this picture:

Three tributaries of causal inference

I worry that because this material has so often been trapped inside elite, mostly American, mostly economics departments, many people feel left out. They may have felt as though their own human capital in statistics was outdated when it absolutely was not. Anything true is beautiful and good. It’s just that this material was a fusion born out of shoeleather and statistics. It was a preoccupation with credibility in the way that a particular group of labor economists—not even economists, not even labor economists, but this particular tradition of labor economists—thought about that, and their own historical context.

The field of causal inference is like the Mississippi River in that its water and creatures came from three distinct tributaries: Princeton Industrial Relations Section, Harvard’s Statistics and Economics, and then everything else that was neither of those, which I simply call “Much More Stuff.” Many of us—myself included—come from the Much More Stuff tradition of empirical methods and we can tell sometimes that our training was different, but perhaps can’t quite put our finger on it.

I have not attempted to write the definitive work on causal inference and do not envy, whatsoever, whichever one of you feels called to do that. It sounds like a thankless and impossible task because, as I said at the start, there’s almost certainly no way to do it without assuming the other rivers do not exist. This book was always a mixtape about a single part of the story of causal inference that I knew, first hand, was simply to many, many of us not accessible. It’s not accessible perhaps because we come from the Much More Stuff rivers.

We did not go to Princeton in the 1970s, 1980s, and 1990s. We did not go to Harvard to get our PhDs in statistics. We didn’t go to Stanford, Berkeley, MIT, or Harvard for economics PhDs—all direct descendants of this Princeton–Harvard fusion. And many of us didn’t even attend the schools where those graduates later taught.

That is not a bad thing—not at all. All of us went where we went, did what we did, because it was exactly what was supposed to happen. But the point is that this approach to empirical work is an odd fusion of things that may not have been the rivers that we were in. For historical reasons, these two rivers swelled like floodwater surges, and at times felt like they had completely taken over the rivers that we were in. And, if you could not see all of it in space and time, it might have even been rough and confusing, dizzying even, tossing us about like a raft on white waters.

Rivers do not travel in straight lines. They meander. Sometimes they meander like a writhing snake, flipping in circles, almost chaotic to watch. Maybe you were like me, just close enough that you watched the river curve right around you. As I said in an earlier chapter, the city of New Orleans, Louisiana, where I lived from 1999 to 2002, is called the Crescent City precisely because the Mississippi River wraps around it in the shape of a crescent. But not everybody lives in the crescent.

Sometimes the meanders can be so extreme that the river breaks off and forms a lake. Some of us maybe feel like we were in one of those lakes. In southeastern Arkansas, there is a lake created centuries ago when a large meander of the Mississippi was cut off from the main channel. Over time, sediment filled in the connecting ends, isolating the loop and forming the lake called an oxbow lake. Oxbow lakes are like the alchemy of living rivers. They were once a part of that river, but over time, as the river gradually changed course, they were left behind.

These are simply things that rivers do, and just like these are things that rivers do, these are also things that science does, and these are things that have been done in causal inference. Causal inference has its own oxbow lakes. It has its own alchemy, just like chemistry does, just like everything does. Humans have been trying to understand a very complex world using anything at their disposal ever since we crawled out of the sea. And, every time we moved ahead, we left things behind that were no longer useful, but which showed we’d been there, just like those oxbow lakes.

I can’t say I lose any sleep at night thinking about oxbow lakes, though, and their tragic story of having been left behind. But if I’m honest, I do lose sleep over the thought that there are people who feel left behind, and I have for a very long time. I am, after all, a teacher. Teachers are by design and calling people who enjoy sharing with others the things that helped them learn, that they think are beautiful, that they think are true. Every teacher is making mixtapes for their students. It’s what we do. We are all doodling on little pieces of paper hearts and stars and rivers taped to cassette tapes filled with a couple dozen songs selected just for them. We give them to one another because we like how it feels to connect people to beautiful, special things.

If this book has helped you, if you liked it, that’s wonderful. I’m glad. I wrote this book for people like me who have papers burning holes inside their hearts and minds—people who feel that if they don’t get those papers out, if they don’t answer their own burning questions, then they’ll burst. I hope that this book will help you write those papers. I hope that this book will help you read other books, other papers, and learn from other people.